Skip to content

docs(AGENT-BEST-PRACTICES): land session-closure rule as small operational artifact (cross-AI converged stance) #712

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
AceHack merged 2 commits into from
Apr 29, 2026
Merged

docs(AGENT-BEST-PRACTICES): land session-closure rule as small operational artifact (cross-AI converged stance) #712

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
102d0f8
docs(AGENT-BEST-PRACTICES): land session-closure rule as small operat…
AceHack Apr 29, 2026
cf57967
review-thread fixes: 3 P1 internal-consistency findings on PR #712 (C…
AceHack Apr 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 22 additions & 0 deletions docs/AGENT-BEST-PRACTICES.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -348,6 +348,28 @@ BP drift.
under the documentation-shepherd's lane in
`docs/BACKLOG.md`.

- **Session-closure rule — after a big round lands, do
not expand; close the round, then test the smallest
bridge.** When a round closes by landing 10+ PRs, a
promoted bead, or a research artifact, the agent does
not extend the round with new conceptual substrate. The
next directed work is either (a) verifying durable-home
Comment thread
AceHack marked this conversation as resolved.
Outdated
landing for the round's substrate, or (b) running the
smallest prototype the round defined. Adding more theory
before the prototype runs uses theory as a substitute
for evidence. The discipline applies recursively —
including to the logging mechanism itself: tick-history
rows are appended only on material state transitions
(PR merged / failed / blocker changed / thread resolved
/ new substrate landed / explicit human ask), not on
Comment thread
AceHack marked this conversation as resolved.
Outdated
pure thread-fix maintenance already recorded in PR
commit messages. Catcher attribution + lineage live on
history surfaces (`memory/feedback_*.md`,
Comment thread
AceHack marked this conversation as resolved.
Outdated
`docs/hygiene-history/`, `docs/research/aurora-immune-governance-bridge-minimal-2026-04-28.md`).
Operational rule, not theoretical framing — the rule's
job is to gate scope-expansion at round-close, not to
explain it.

---

## How rules become stable
Expand Down
Loading